AI Sameness: Why Imperfection Is Becoming a Competitive Advantage

A line from the current design debate captures something I've been noticing in image production for a while: AI smooths everything, making it faster, more efficient, more comparable. That "more comparable" is exactly the problem. When every studio has access to the same models, the look converges – brands that used to be distinguishable start to blend together. The industry has started calling this "AI sameness," and the counter-movement against it is becoming just as noticeable as the AI wave itself.

What AI sameness actually means

This isn't about one bad image, it's a pattern: interchangeable interfaces, campaign visuals, and brand worlds that are technically flawless and still feel soulless. The reason lies in the nature of optimization itself. A model trained for maximum plausibility produces, when in doubt, the most probable output, not the most distinctive one. That's technically nothing new in CGI – we've known this problem for years from pure rendering aesthetics, long before generative AI entered the picture. What's new is the speed at which this smoothing is now spreading across the entire industry, because suddenly everyone has access to comparably good tools.

Why imperfection is becoming valuable again

The counter-movement currently being described in the design debate doesn't surprise me: deliberately imperfect elements, handmade textures, sketchy typography, real materiality instead of digital polish. This isn't nostalgic retreat, it's an economically sound reaction. When perfection becomes a commodity, it loses its value as a point of differentiation. I've actually described this in practice before, just from a different angle: photography has an advantage over CGI that's rarely named – chance. An unexpected fall of light, an unplanned gesture. CGI is perfect because it's controlled. That very control is now becoming the problem, once every studio has the same control over the same tools.

The difference between imperfection and carelessness

This is where the current debate falls short if adopted uncritically. "More imperfection" isn't a free pass for sloppy work, and a deliberately scrawled logotype is just as much a design decision as a clean interface – just one that consciously breaks the standard. That requires knowing the standard precisely. Deliberate imperfection only works when it's part of a clearly defined system, not its absence. That's exactly what was at stake when we defined a visual language as a system rather than a matter of taste at Kampart: a rulebook that also specifies where to deliberately break the rules is something different from having no rulebook at all.

What this means for our production

In practice, this doesn't mean scaling back AI tools for us, but deciding more precisely where to use them. Variant creation, cleanup, retouching – areas we've described in content production – benefit from speed without the character of an image depending on it. The actual creative differentiator, the point where one image or campaign sets itself apart from the next, still comes from human decisions: which detail stays deliberately imperfect, which rule gets broken, where photography works better than rendering. That's also why physical experiences are gaining importance right now instead of disappearing – something I've described in more detail elsewhere: the more digital and polished the world becomes, the more value accrues to what can't be smoothed over.

Conclusion

AI sameness isn't a footnote in the design debate, it's a direct consequence of good tools now being available to everyone. Whoever relies purely on the technology gets exactly the result everyone else gets: comparable, clean, interchangeable. The competitive advantage now sits exactly where it always did, just hidden underneath perfection – in the deliberate decision of where to deviate from the rule, and in the courage to take responsibility for it.